Scientists often test drugs in mice. Now some cancer patients are doing the same - with the hope of curing their own disease.

They are paying a private lab to breed mice that carry bits of their own tumors so treatments can be tried first on the customized rodents. The idea is to see which drugs might work best on an individual's specific cancer.

The mice may help patients make what can be very hard choices under difficult circumstances. Studies can suggest a certain chemotherapy may help, but patients wonder whether it will work for them. Often there's more than one choice, and if the first one fails, a patient may be too sick to try another.

"What I'm doing is personalized cancer treatment. It's the wave of the future," said Eileen Youtie, a Miami woman using mice to guide care for her hard-to-treat form of breast cancer. "Part of this is trying to eliminate chemos that are not going to work on me. I don't want to waste time taking them and poison my body."

There are no guarantees the mice will help.

But there are some early encouraging reports, said Dr. Len Lichtenfeld, deputy chief medical officer of the American Cancer Society.

One study of 70 patients found the mice generally reflected how well patients responded to various drugs. But there is no evidence that using mice is any better than care based on medical guidelines or the gene tests that many patients get now to help pick drugs.

Mouse testing costs $10,000 or more, and insurers don't cover it. It takes several months, so patients usually have to start therapy before mouse results are in.

Youtie spent $30,000 "because I want them to test all the possible drugs," even some for other types of cancer.

That approach helped Yaron Panov, a 59-year-old Toronto man diagnosed four years ago with liposarcoma, a soft-tissue cancer. No specific drugs were recommended, and "I was given six months to live," he said.

Tests on his avatar mice suggested the first drug he was prescribed would not work but that one for colon cancer might.

"It was working on the mice so I knew it would work on me," he said. "It's such a boost of confidence" and it makes it easier to endure side effects, said Panov, whose cancer is in remission.

Reuven Moser, a 71-year-old man from Tel Aviv said his avatar mice confirmed that drugs prescribed for colon cancer that had spread to his liver were a good option.

"Most of the time the oncologists want to follow a protocol, but they don't know how it will affect the patient," Moser said. "It was very reassuring" to see the mice respond, he said. Moser's mice were bred in February and he is still undergoing treatment.

Dr. Andrew Gaya of Leaders in Oncology Care, a private clinic in London, helped lead the 70-patient study of avatar mice and gave the results at a cancer conference in September. It looked back at how well mice performed in patients whose outcomes from treatment were already known. About 70 percent of the time, tests in the mice suggested something that turned out to have helped the patients, he said. And if something had not worked in the mice it almost never worked in a patient.